I recently upgraded to CUDA 2.0, and now when I compile my old programs I get:
“Error: Unaligned memory accesses not supported”.
What is causing this, and how can I fix it? (is there an __align macro?)
This is the offending piece of code:
struct rc4_state {
u_char perm[256];
u_char index1;
u_char index2;
};
... *snip* ...
outbuf[i] = inbuf[i] ^ state->perm[j];